A highly efficient synthesis of fibrinogen receptor antagonist L-734,217 via a novel chemoselective silyl-mediated conjugate addition of δ-lactams to 4-vinylpyridine

Article abstract of DOI:10.1021/jo951214f

A highly practical chromatography-free six-step synthesis of L-734,217 suitable for large scale preparation is described. The key chiral pyridine acid intermediate (R)-1 was prepared in four steps based on a novel chemoselective silyl-mediated conjugate a

Full text of DOI:10.1021/jo951214f

J . Org. Chem. 1996, 61, 215-222
215
A High ly Efficien t Syn th esis of F ibr in ogen Recep tor An ta gon ist
L-734,217 via a Novel Ch em oselective Silyl-Med ia ted Con ju ga te
Ad d ition of δ-La cta m s to 4-Vin ylp yr id in e
J ohn Y. L. Chung,* David L. Hughes, Dalian Zhao, Zhiguo Song, David J . Mathre,
Guo-J ie Ho, J ames M. McNamara, Alan W. Douglas, R. A. Reamer, Fuh-Rong Tsay,
Richard Varsolona,† J ames McCauley,† Edward J . J . Grabowski, and Paul J . Reider
Department of Process Research and Department of Analytical Research, Merck Research Laboratories,
Merck & Co., Inc., P. O. Box 2000, Rahway, New J ersey 07065
Received J uly 6, 1995X
A highly practical chromatography-free six-step synthesis of L-734,217 suitable for large scale
preparation is described. The key chiral pyridine acid intermediate (R)-1 was prepared in four
steps based on a novel chemoselective silyl-mediated conjugate addition of ethyl (2-oxopiperidin-
1-yl)acetate to 4-vinylpyridine and a highly productive, recyclable, kinetic resolution with quinine.
Subsequent salt breaking/peptide coupling with benzyl 3-(R)-aminobutyrate (2) in a biphasic system,
followed by concomitant hydrogenation of the pyridine ring and debenzylation afforded L-734,217
in 20% overall yield (30% with one recyle) from 2-piperidone. The mechanism of this key conjugate
addition to 4-vinylpyridine was studied by 13C NMR.
In tr od u ction
ate (R)-1 was synthesized based on a novel chemoselec-
tive silyl-mediated conjugate addition of a 2-piperidone
derivative to 4-vinylpyridine and a highly productive,
recyclable, kinetic resolution with quinine. These new
developments render alternate methods of asymmetric
synthesis unnecessary.
Vascular occlusion by thrombus formation has been
implicated as a primary cause of a variety of cerebral and
cardiovascular diseases that affect over one million
patients within the United States each year.1 In search
of pharmacological inhibitors of these events, a strategy
based on the inhibition of plasma fibrinogen binding with
the platelet GPIIb/IIIa complex has been demonstrated
to have wide-spectrum platelet inhibitory activity without
side effects associated with other approaches.2,3 Thus,
extensive efforts have been directed toward the discovery
and synthesis of such therapeutic agents.3,4 L-734,217,
a non-peptide, orally active fibrinogen receptor antago-
nist, was the result of such a search and is currently
undergoing clinical studies.5
The initial synthetic route to this compound proceeded
from 4-piperidineethanol via a 14-step sequence.6 How-
ever, a more concise and practical route was desired.
Herein we disclose novel chemistry leading to an efficient,
highly practical six-step synthesis of L-734,217 that
requires no chromatography and uses readily available
starting materials (Scheme 5). The key chiral intermedi-
Our retrosynthetic analysis is outlined in Scheme 1.
We envisioned that a pyridylethylation reaction with a
2-piperidone derivative using 4-vinylpyridine would offer
quick access to the racemic pyridine acid 1, which after
an efficient resolution should offer a practical route to
(R)-1. The employment of 4-vinylpyridine not only
resolved the availability issue of starting material but
also significantly improved the efficiency of the route by
having the pyridine moeity acting as a piperidine sur-
rogate thereby avoiding the use of an expensive nitrogen
protecting group. After peptide coupling of (R)-1 with
the commercially available chiral benzyl â-aminobutyrate
2,7,8 concomitant hydrogenation of pyridine to piperidine
and debenzylation should afford L-734,217.
Resu lts a n d Discu ssion
P r ep a r a tion of Ra cem ic P yr id in e Acid 1. Our
initial approach to racemic 1 was to effect a classical
conjugate addition9 of 3-carbethoxy-2-piperidone (3) to
4-vinylpyridine followed by hydrolysis-decarboxylation
of the carbethoxy group (Scheme 2). The conjugate
addition of 3 to 4-vinylpyridine was carried out in
refluxing ethanol with catalytic sodium ethoxide in the
presence of a trace amount of hydroquinone. Without
isolation, the resulting crude adduct 4 was directly
saponified and decarboxylated to afford pure compound
5 in 77% overall yield after crystallization. Subsequent
N-alkylation with ethyl bromoacetate followed by saponi-
fication of the resulting ester 11 afforded the pure
racemic pyridine acid 1 in 70% overall yield. Despite the
† Department of Analytical Research.
X Abstract published in Advance ACS Abstracts, November 15, 1995.
(1) (a) Nichols, A. J .; Ruffolo J r., R. R.; Huffman, W. F.; Poste, G.;
Samanen, J . Trends Pharmacol. Sci. 1992, 13, 413. (b) Coller, B. S. N.
Engl. J . Med. 1990, 322, 248. (c) Ellis, S. G.; Bates, E. R.; Schaible, T.;
Weisman, H. F.; Pitt, B.; Topol, E. J . J . Am. Coll. Cardiology 1991,
17, 89b.
(2) (a) Kleinman, N. S.; Ohman, E. M.; Keriakes, D. J .; Ellis, S. G.;
Weisman, H. F.; Topol, E. J . Circulation 1991, 84 (Suppl. II), 522. (b)
Aiken, J . W.; Gorman, R. R.; Shebuski, R. J . Prostaglandins 1979, 17,
483.
(3) Blackburn, B. K.; Gadek, T. R. In Annual Reports in Medicinal
Chemistry; Bristol, J . A., Ed.; Academic Press, Inc.: San Diego, 1993;
Vol. 28, Chapter 10, p 89.
(4) Cheng, S.; Craig, W. S.; Mullen, D.; Tschopp, J . F.; Dixon, D.;
Pierschbacher, M. D. J . Med. Chem. 1994, 37, 1 and references therein.
(5) (a) Duggan, M. E.; Naylor-Olsen, A. M.; Perkins, J . J .; Anderson,
P. S.; Chang, C. T.-C.; Cook, J . J .; Gould, R. J .; Ihle, N. C.; Hartman,
G. D.; Lynch, J . J .; Lynch, R. J .; Manno, P. D.; Schaffer, L. W.; Smith,
R. L. J . Med. Chem. 1995, in press. (b) Duggan, M. E.; Egbertson, M.
S.; Ihle, N.; Hartman, G. D.; Turchi, L. M.; Hoffman, W. F. U.S. Patent
5,281,585, J an 5, 1994.
(7) Chiral benzyl â-aminobutyrate 2 is commericially available from
Celgene Corp., 7 Powder Horn Drive, Warren, NJ 07059.
(8) For a review on enantioselective synthesis of â-amino acids,
see: J uaristi, E.; Quintana, D.; Escalante, J . Aldrichim. Acta 1994,
27, 3.
(9) Siperko, L. G.; Smith, F. X. Synth. Commun. 1979, 9, 383, and
references therein.
(6) The initial synthesis of L-734,2175 required 17 steps (14 linear
steps and 3 steps to prepare the tert-butyl ester analog of 2), 11
chromatographies, and several raw materials and reagents that are
not readily available in bulk.
0022-3263/96/1961-0215$12.00/0 © 1996 American Chemical Society

216 J . Org. Chem., Vol. 61, No. 1, 1996
Chung et al.
Sch em e 1
Sch em e 3
°C to afford pyridine-piperidone 5 in 75-80% assay yield
after aqueous quench. Again, due to the high hydrophilic
nature of 5, the pure product was isolated in only 50%
yield after extractive workup and purification by silica
gel filtration.
One simple solution to this problem was to avoid 5 as
an intermediate by reversing the reaction sequence of the
conjugate addition and the alkylation step, which should
provide the lipophilic compound 11 from piperidone-ester
7. However, the chemoselectivity of the conjugate addi-
tion of 7 to 4-vinylpyridine was in question, since one
would expect the addition to take place primarily at the
ester R-carbon due to its higher acidity over the lactam
R-carbon. Nevertheless, 2-piperidone was first N-alky-
lated with ethyl bromoacetate, and the resulting piperi-
done-ester 7 (84%) was reacted with 4-vinylpyridine
under the above described conditions (2.2 equiv of silyl
reagent). Pleasingly, a 65-70% assay yield of crude 11
was obtained after aqueous workup with negligible loss
to the aqueous phase. The main byproducts were the bis-
pyridyl adduct 13 and starting ester 7 at 15-20% and
10-15%, respectively. Surprisingly, none of the ester
adduct was detected. The addition of 4-vinylpyridine
appears to be completely selective toward the lactam
R-carbon over the ester R-carbon. To our best knowledge,
the silyl-mediated conjugate addition of 4-vinylpyridine
and the observed chemoselectivity are unprecedented.
This interesting chemoselectivity is contrary to the pKa
values of carbonyl groups where one would expect that
enolization of the ester would precede enolization of the
lactam.13
Sch em e 2
success of this approach, it was not suitable for large scale
use due to the hydrophilic nature of compound 5 which
made its isolation rather cumbersome. In addition,
compound 3 was not available in commercial quantities.
Its preparation via Raney nickel reduction of the di-
methyl malonate-acrylonitrile Michael adduct proved to
be erratic.10 This prompted us to search for a more direct
means to prepare 1 without using 3.
The obvious choice was to use the readily available
2-piperidone (6) as the starting material. However,
initial attempts to prepare 5 by adding 4-vinylpyridine
to the dianion of 611 gave no reaction.12 Reasoning that
the hardness of the dianion was not compatible with
4-vinylpyridine, we then surmised that the “softer” N,O-
disilyl ketene 9 should react with 4-vinylpyridine, espe-
cially if 4-vinylpyridine is further activated by N-silyla-
tion (Scheme 3). Thus, treatment of 2-piperidone with
2.2 equiv each of TMSOTf and Et3N or 2.2 equiv each of
TMSCl, Et3N, and NaI in acetonitrile at 0-22 °C gave
the presumed silyl ketene N,O-acetal 9, which without
isolation was directly treated with 4-vinylpyridine at 0
This reversal of expected chemoselectivity from ester
to lactam in the silyl-mediated reaction is most likely due
(13) Further support for this contra-thermodynamic phenomenon
was that when
7 was treated with 9-BBN-OTf (1.05 equiv) and
triethylamine (1.1 equiv) in CCl4 it gave only the C-boronate of the
ester B, presumably via the ester O-boronate.14 None of the products
derived from boronation of the lactam were detected.15
(10) (a) Koelsch, C. F. J . Am. Chem. Soc. 1943, 65, 2458. (b)
Liebman, A. A.; Malarek, D. H.; Dorsky, A. M.; Kaegi, H. H. J .
Heterocycl. Chem. 1974, 11, 1105.
(11) For preparation of the dianion of 2-piperidone, see: Yamamoto,
Y.; Morita, Y. Chem. Pharm. Bull. 1984, 32, 2555.
(12) (a) Conjugate addition of N,N-dibutylpropionamide to 2-vi-
nylpyridine in DMF at 110 °C with catalytic potassium tert-butoxide
in 51% yield has been reported: Hintermeier, K.; Brachel, H. Ger.
Offen. 2025009, Dec 9, 1971. (b) For various Michael reactions of amide
and lactam enolates with enones, see: Heathcock, C. H.; Henderson,
M. A.; Oare, D. A; Sanner, M. A. J . Org. Chem. 1985, 50, 3019.
B
(14) Irreversible 1,3-boryl shifts are known: Paetzold, P.; Biermann,
H.-P. Chem. Ber. 1977, 110, 3678.
(15) Brown, H. C.; Dhar, R. K.; Ganesan, K.; Singaram, B. J . Org.
Chem. 1992, 57, 499.

Synthesis of Fibrinogen Receptor Antagonist L-734,217
Sch em e 4
J . Org. Chem., Vol. 61, No. 1, 1996 217
to the higher nucleophilicity16 of the lactam toward
trimethylsilyl triflate and the formation of the more
stable imidate cation relative to the ester counterparts.
Thus, the following deprotonation and alkylation with
N-silyl-4-vinylpyridine (A) should go to the lactam R-C
instead of that of the ester. This explanation seems
consistent with 13C NMR studies wherein O-silyl imidate
cations 17 and 18 were observed17 (see Figure 1 in the
supporting information for C13 chemical shift assign-
ments), and no intermediates resulting from silylation
of the ester group were detected.
The proposed mechanism for this reaction is outlined
in Scheme 4. 13C NMR examination of a mixture of
piperidone ester 7 and TMS-OTf (1.2 equiv) in CD3CN
at 0 °C showed that lactam ester 7 was completely
converted to imidate cation 17. Upon addition of tri-
ethylamine, competition between lactam 7 and triethyl-
amine for available silyl groups was indicated by line
broadening in triethylamine and shift averaging in
lactam 7 with no observable evidence for intermediate
10. However, the formation of C,O-bis-silylated 18 was
observed when CD2Cl2 was substituted for CD3CN. In
CD3CN at -20 °C, the R-carbon signal of 17 was notably
broadened, possibly due to silyl exchange or to deuterium
incorporation from the solvent. In addition, there was
evidence of accompanying proton incorporation in the
solvent spectral pattern. These observations suggested
the formation of intermediate 10. It appeared that rapid
trapping of 10 by electrophiles (TMS-OTf in this case)
present in the system could have precluded it from being
detected.18 Subsequent incremental addition of 4-vi-
nylpyridine at 0 °C resulted in an averaged spectra of
21 and 11, indicating they are in dynamic equilibrium.
Presumably, the conjugate addition of 10 to A proceeded
via 19 and 20, which was then protonated to give 21 or
alternatively underwent another conjugate addition with
A to give the bis-adduct 13 side product. At the end of
the reaction, as all 21 was converted to 11, almost all
triethylamine was silylated (as R3N+SiMe3).
Differences between the two solvents used in the NMR
study are of importance. A moderating effect of aceto-
nitrile19 upon the silylating activity of TMS-OTf may be
responsible for the finite lifetime of silylated 4-vinylpyr-
idine in that solvent. Such an effect could explain the
relatively poor reaction yield in methylene chloride
especially when (i-Pr)2NEt (Hunig’s base) was used, both
of which have no moderating effect leading to rapid
polymerization of 4-vinylpyridine and, thus, little of the
desired product.
In an effort to improve the yield, we examined alterna-
tive solvents, bases, silylating agents, temperature, and
stoichiometry. As mentioned above, a significant solvent
effect was observed, with acetonitrile being the best
solvent. Methylene chloride is also a good solvent;
however, it only works well with sterically less hindered
tertiary amine bases. The yield deteriorates with tolu-
ene, isopropyl acetate (IPAC), or ethereal solvents as the
cosolvent with acetonitrile. Et3N, Me2EtN, and (Me2-
NCH2)2 (TMEDA) all gave similar yield in acetonitrile,
whereas Hunig’s base almost completely shut down the
reaction especially in dichloromethane. It appeared that
small tertiary bases were needed to moderate the sily-
lation activity20 to prevent rapid polymerization of 4-vi-
nylpyridine. Hindered bases such as Hunig’s base could
not be silylated and therefore have no moderating effect.
(16) Carbonyl oxygen of amides (lactam here) is more basic than
that of esters. This is shown by their pKa values (Data from: Organic
Chemistry; 3rd ed.; Streitwieser, A.; Heathcock, C. H., Eds.; Macmillan
Publishing Co.: New York, 1985):
(18) For iodotrimethylsilane-mediated 2-monohalogenation of 4-aza-
5R-androstan-3-one steroids, see: King, A. O.; Anderson, R. K.;
Shuman, R. F.; Karady, S.; Abramson, N. L.; Douglas, A. W. J . Org.
Chem. 1993, 58, 3384.
(19) For silylation of acetonitrile by TMS-X, see: (a) Emde, H.;
Simchen, G. Synthesis 1977, 636. (b) Kira, M.; Hino, T.; Sakurai, H.
J . Am. Chem. Soc. 1992, 114, 6697, footnote 6.
(17) For formation of O-silyl imidate cations, see: (a) Frick, U.;
Simchen, G. Liebigs Ann. Chem. 1987, 839. (b) Simchen, G.; Kober,
W. Synthesis 1976, 259.
(20) For silylation of tertiary amines, see ref 21, pp 14-15.

218 J . Org. Chem., Vol. 61, No. 1, 1996
Chung et al.
Ta ble 1. Th er m od yn a m ic Solu bility of Dia ster eom er ic
Sa lts 20 a n d 21
Weaker bases such as pyridine do not work well, probably
because they are not basic enough to deprotonate inter-
mediate 17. With regard to silylating agents,21 TMSOTf
and TMSCl/NaI performed with similar efficiency, whereas
bis(trimethylsilyl)trifluoroacetamide/cat. TfOH and TMS-
Br gave a slower rate of reaction. We found bulky
silylating agents such as triethylsilyl, triphenylsilyl, and
tert-butyldimethylsilyl triflate and also their iodides
somewhat suppress the formation of bis-adduct 13 to a
level of 10-15% and improve the yield of 11 to 75-80%.
Bulky silyl triflates, in general, gave a 5-8% better yield
than the corresponding iodides. On the basis of cost and
availability, triethylsilyl chloride (TES-Cl)/NaI was cho-
sen for further development.
We found slow addition of 4-vinylpyridine was neces-
sary to minimize the formation of polymerization prod-
ucts. Typically an addition time of 2-3 h was used
(0.01-10 mol scale). Attempts to drive the reaction to
completion by charging more 4-vinylpyridine did not
improve the yield. It appeared that 4-vinylpyridine
reacted with both starting piperidone-ester and the
product in such a way that there was no net increase of
the product. The net result was increased formation of
the polypyridyl adduct impurities. Attempts to minimize
the formation of bis-adduct 13 by having excess triethyl-
ammonium triflate in the reaction, hoping to favor the
protonation pathway of 20, however, did not alter the
product-impurity ratio. Also, in theory, the silylating
agent is a catalyst; however, when 0.25 equiv of TESCl
was used the yield dropped to 50-60%.
The current optimized procedure involves addition of
TES-Cl (1.1 equiv) to a 1 M acetonitrile solution of lactam
7, NaI (1.1 equiv), and Et3N (1.2 equiv) at 0 °C. After
aging at 20 °C for 1 h, 4-vinylpyridine (1.15 equiv) was
added over 2 h at 0 °C (0.1 mol), followed by aging for 2
h at the same temperature. The assayed yield at this
point was 75%. After extractive workup, the crude
pyridine-ester 11 was obtained in 70% assay yield as an
2-propanol (IPA) solution along with two impurities, 7
(8%), and bis-adduct 13 (9%) after a solvent switch. This
mixture was used directly in the next step.
Due to the high water-solubility of 1, the hydrolysis of
11 was carried out under near water-free condition to
facilitate the isolation. The optimized procedure involves
treating the above 2-propanolic solution of 11 with 50%
aqueous NaOH (1.3 eq.). After pH adjustment with
concentrated HCl to its isoelectric point (pH 5.5) and
removal of NaCl by filtration, the filtrate was azeotro-
pically dried and treated with hexane to crystallize the
racemic pyridine acid 1 in 91% yield with 97% purity.
The two acid impurities, 8 and 14, derived from esters 7
and 13 were completely rejected in the crystallization.
Resolu tion of Ra cem ic P yr id in e Acid 1. With a
practical synthesis of racemic pyridine acid 1 in hand,
we next sought an efficient resolution.22 Initial studies
of the chiral phenethylamine diastereomeric salts of
racemic 1 indicated that resolution by selective crystal-
lization was possible; however, this material required
four recrystallizations from IPA to achieve 96%+ ee in
only 32% yield (maximum 50%).
solubility of 20
(mg/mL of IPA)
solubility of 21
(mg/mL of IPA)
temp (°C)
5.5
ratio (21/20)
14.4
63.8
4.4
Ta ble 2. Resolu tion Ba sed on Kin etics of Cr ysta l
Gr ow th
concn of 20
concn of 21
temp (°C)
0
(mg/mL of IPA)
(mg/mL of IPA)
ratio (21/20)
11
162
14.7
alkaloids, cinchonine, quinine, and strychnine, gave crys-
talline solids, whereas cinchonidine and quinidine gave
oils. Quinine was chosen for further development based
on solubility diffences between diastereomeric salts and
cost, availability, and toxicity considerations. The reso-
lution was carried out at a concentration of 3 mL of IPA/g
of the salts at 0 °C, which gave 96-98% ee after crystal-
lization and 42-46 wt % recovery. Thermodynamic
solubility data have been obtained on both the diaster-
eomeric salts 20 and 21 (Table 1). The data indicated
that the solubility difference between the diastereomeric
quinine salt is only 4.4 fold at 5.5 °C. At the concentra-
tions used in the resolution (428 mg/g solvent), this would
mean at equilibrium, the product crystallized would be
57% ee. However, the growth kinetic of the desired
diastereomeric salt 20 is much faster than the undesired
diastereomeric salt 21. Data in Table 2 show that after
a 14 h age at 0 °C, the supernatant contains 11 mg/mL
of 20 and 162 mg/mL of 21, which is a 14.7 fold difference
between the two salts. These results indicated that the
resolution is a kinetic phenomenon. On the basis of
thermogravimetric and NMR studies, this large differ-
ence in the growth rates was subsequently determined
to be related to the fact that the desired (R) diastereo-
meric salt forms an IPA solvate whereas the (S) diaster-
eomeric salt does not. Further studies revealed that this
kinetic ratio was stable at 0-10 °C for at least 15 h. Thus,
there was sufficient time for isolation without disturbing
the diastereomeric ratio. Also, it was found that the
presence of water was deleterious due to the formation
of water solvates which gave no resolution.
Based on these results, a reliable resolution procedure
was developed. Pyridine acid 1, quinine, and dry IPA (1
g/3 mL) were combined and then heated at 65 °C for 15
min. The mixture was seeded with >99% ee quinine salt
at 45 °C and then aged at 22 °C for 15 h. The mixture
was cooled to 10 °C within 30-45 min, and then the solid
was filtered. After washing with 3-4 bed volumes of cold
THF/hexane (1:1), the solid was vaccum dried. Since the
(R) diastereomeric salt forms an IPA solvate, the final
solid contain ca. 9% IPA.
Purity of the resolved pyridine acid quinine salt was
89.5 wt % (40 wt % for pyridine acid alone) with 9.3 wt
% of IPA and 1.2 wt % of NaCl. The optical purity of the
chiral acid was determined to be 98% ee via the (R)-
naphthylethylamine derivatives on a normal phase HPLC
column (Zorbax SIL) using CH2Cl2/IPA/NH4OH (980/17/
3) as the mobile phase. The (R)-naphthylethylamine
derivatives was prepared directly from the quinine salt
without having to break salt first, requiring only 10-15
min.
Subsequently, five alkaloids were examined for their
ability to form a crystalline solid with pure (R)-1. Three
Recyclin g of th e Moth er Liqu or . To further im-
prove the efficiency, we found that the undesired (S)
diastereomeric salt in the resolution mother liquor could
be directly racemized without breaking the salt by
(21) For a review, see: Emde, H.; Domsch, D.; Feger, H.; Frick, U.;
Gotz, A.; Hergott, H. H.; Hofmann, K.; Kober, W.; Krageloh, K.;
Oesterle, T.; Steppan, W.; West, W.; Simchen, G. Synthesis 1982, 1.
(22) Enantiomers, Racemates, and Resolutions, J acques, J .; Collet,
A.; Wilen, S. H.; Eds.; J ohn Wiley & Sons: New York, 1981.

Synthesis of Fibrinogen Receptor Antagonist L-734,217
J . Org. Chem., Vol. 61, No. 1, 1996 219
Sch em e 5. Six-Step Syn th esis of L-734,217
treatment with 2.0 equiv of NaOEt at 85 °C for 3-5 h.
Preliminary results indicated that racemization does not
take place with 1.2 equiv of NaOEt. After neutralization
with concentrated HCl to pH 8.0-8.4 and dilution with
THF, the NaCl was filtered. The filtrate was azeotropi-
cally dried and reconstituted to a concentration of 1 g of
quinine salt per 3 mL of IPA. This mixture was resolved
as described above to give the desired pyridine acid
quinine salt in 36% yield and 97+% ee. THF was added
to reduce the viscosity of IPA and to facilitate and
filtration. It also decreased the solubility of NaCl in the
solvents.
It is interesting to note that in the hydrogenation step,
lowering the hydrogenation temperature to 20 °C pro-
vided selective debenzylation of 15 without reduction of
the pyridine ring. By carefully monitoring the reaction
and terminating the reaction after uptake of 1 equiv of
hydrogen, product 23 was isolated in 86% yield after
crystallization from a mixture of CH3CN-MTBE (1:2).
An attempt to directly racemize the quinine salt by
heating the mother liqour in a sealed tube at 150 °C gave
no racemization. About 20% of the pyridine acid decom-
posed.
Sa lt Br ea k in g, P ep tid e Cou p lin g, a n d Hyd r oge-
n a tion . With the chiral pyridine acid quinine salt in
hand, the salt was partitioned between aqueous sodium
hydroxide and IPAC or tert-butyl methyl ether (MTBE).
After two extractions with IPAC, less than 1 mol %
quinine was left in the aqueous layer.
This aqueous solution of pyridine acid (R)-1 was then
used directly in the 1-[3-(dimethylamino)propyl]ethyl-
carbodiimide hydrochloride (EDC)-mediated coupling
with benzyl 3(R)-aminobutyrate (2) in the presence of
IPAC as the second phase. In the absence of this organic
phase, significant polyamidation of the product ester 22
by amine 2 took placed. This organic layer efficiently
removed the product ester 22 from amine 2 as it formed
in the aqueous layer, thus eliminating the amidation side
reaction. After extractive workup of the reaction, the
coupled product 22, together with trace amount of the
quinine left from the last step, was turned over to
methanol and then hydrogenated over 10% palladium on
carbon at 50 °C under 40 psi hydrogen. After the catalyst
is removed, L-734,217 is then crystallized from a mixture
of MeOH-CH3CN. All the reduced quinine is rejected
in this crystallization according to HPLC. Starting from
98% ee resolved pyridine acid and >99% ee benzyl 3(R)-
aminobutyrate (2), the final product is >99% de and the
overall yield is ca. 85% based on quinine salt 20.
One remaining concern in the coupling reaction is the
use of the expensive reagent EDC. Preliminary work
with DCC showed that the major difficulty is dissolution
of the two starting materials, pyridine acid, and 3-ami-
nobutyrate, into a solvent that will also dissolve DCC.
Toward that end, it was found that a 1:1 mixture of
THF-H2O is a good solvent for the coupling reaction, and
dicyclohexylurea is removed from the product solution
in IPAC as a solid. The yield was nearly quantitative
and subsequent hydrogenation gives L-734,217 with no
significant difference in impurity profile. Thus, DCC is
used as successfully as EDC.
In summary, a highly practical chromatography-free
six-step synthesis of L-734,217 suitable for large scale
preparation was developed (Scheme 5). The key chiral
pyridine acid intermediate (R)-1 was prepared in four
steps based on a novel chemoselective silyl-mediated
conjugate addition of ethyl (2-oxopiperidin-1-yl)acetate
to 4-vinylpyridine and a highly productive, recyclable,
kinetic resolution with quinine. Subsequent salt break-
ing/peptide coupling with benzyl 3(R)-aminobutyrate (2)
in a biphasic system, followed by concomitant hydrogena-
tion of the pyridine ring and debenzylation afforded
L-734,217 in 20% overall yield (30% with one recyle) from
2-piperidone.
In an effort to further improve the synthesis, attempts
to affect asymmetric conjugate addition of piperidone-
ester 5 to 4-vinylpyridine with chiral chlorosilane were
explored. Preliminary studies with dimethyl((-)-men-

220 J . Org. Chem., Vol. 61, No. 1, 1996
Chung et al.
thyloxy)silyl chloride 2423 gave 45% yield of pyridine-ester
7, however, with no asymmetric induction.
pH should be 1.5-2.5. The acidic solution (pH ∼ 2) was
extracted with 1:1 IPAC/hexane (2 × 160 mL). The aqueous
layer was assayed by HPLC showing 24 g of product 11 (76%
yield), 1.2 g of 7 (6%) and 4.7 g of bis-pyridyl impurity 13 (11%).
To the aqueous solution was added IPAC (120 mL), and the
mixture was cooled to 5-10 °C. With vigorous stirring, it was
then basified to pH 9.5-10 by the slow addition of solid sodium
bicarbonate (10 g; to pH 6) and 5 N NaOH (˜22 mL; to pH
9.7). The layers were separated, and the aqueous layer was
extracted with toluene (2 × 150 mL). The combined organic
layers were washed with saturated aqueous sodium bicarbon-
ate (3 × 50 mL). The organic layer was azeotropically dried
by distillation at 60 °C under reduce pressure. After ∼400
mL distilled out, distillation was terminated and 150 mL dry
toluene (total volume ) 200 mL) and 12 g of silica (60-200
mesh) were added. After stirring for 1 h, the mixture was
filtered and the filter cake was washed with 100 mL toluene.
The combined filtrate was assayed to contain 21.9 g (70%
overall yield; 91.8% recovery) of product 11. After distilling
off most of the solvent, the batch was turnovered to an IPA
solution with a final concentration of 25 wt % (86 g) in IPA.
This solution was used as is in the next step. An analytically
pure oily sample of 11 was prepared by flash chromatogra-
Exp er im en ta l Section
Gen er a l. All reactions were conducted under an atmo-
sphere of dry N2. As necessary, most of solvents and reagents
were dried over 3 Å or 4 Å molecular sieves and residual water
was determined by Karl Fisher titration. Melting points were
uncorrected. 1H and 13C NMR spectra were collected at 250
and 63 MHz, respectively, from samples in the specified
deuterated solvent. Assay yields of product were determined
by HPLC analysis using the corresponding pure products as
standards.
P r ep a r a tion of Eth yl (2-Oxop ip er id in -1-yl)a ceta te (7).
A dry 5 L four-necked round bottom flask equipped with a
mechanical stirrer, an addition funnel, nitrogen inlet, cooling
unit, and thermometer probe was charged with 2-piperidone
(160.00 g, 1.61 mol), anhydrous THF (3.36 L), and TMEDA
(206.3 g, 1.78 mol). The solution was cooled to -10 °C, and
n-butyllithium (1.6 M in hexane, 1.06 L, 1.70 mol) was slowly
added over a 60 min period, while keeping the internal
temperature less than 0 °C. After the addition, the reaction
mixture was stirred at 0-5 °C for 1 h. The reaction mixture
was cooled to -10 °C, and ethyl bromoacetate (283.1 g, 1.70
mol) was added over 15 min while maintaining the internal
temperature less than 0 °C. The reaction mixture was stirred
at 0 °C for 15 min and then allowed to warm to 23 °C and
aged at this temperature for a 2 h period. The reaction
mixture was cooled to -5-0 °C and quenched into a solution
of NaCl (170 g) in 2 N HCl (1.78 L), while keeping the internal
temperature less than 20 °C. The resulting aqueous phase
should be at pH 6. The mixture was transferred to a 12 L
separatory funnel and the two layers were separated. The
aqueous layer was extracted with isopropyl acetate (IPAC) (3
× 1 L). The combined organic layers were concentrated to near
dryness and then azeotropically dried with CH3CN (3 × 600
mL) (50 °C, house vacuum). The mixture was filtered to
remove a small amount of NaCl after the azeotropic distilla-
tion. The filter cake was washed with CH3CN (500 mL). The
combined filtrate was assayed by HPLC showing 253.5 g of
product (84% yield; 83-85 LC area % purity at 215 nm) as a
36 wt % solution. The water content was 0.06 mg/mL and
contained <2 mol % of THF and IPAC by NMR. This brown
solution was used as is in the next step. Pure solid product
can be isolated by recrystallization from IPAC/hexanes: mp
1
phy: MS (EI) m/ z 290 (M
+); H NMR (CDCl3) δ 1.09 (t, J )
7.1 Hz, 3H), 1.50 (m, 1H), 1.60-1.90 (m, 2H), 2.04 (m, 1H),
2.20 (m, 1H), 2.54 (m, 2H), 3.10-3.30 (m, 2H), 3.77 (A of AB,
J ) 17.2 Hz, 1H), 4.01 (q, J ) 7.1 Hz, 2H), 4.03 (B of AB, J )
17.2 Hz, 1H), 6.99 (d, J ) 6.0 Hz, 2H), 8.30 (d, J ) 6.0 Hz,
2H); 13C NMR (CDCl3) δ 9.7, 17.3, 22.2, 27.9, 28.0, 36.2, 44.6,
44.9, 56.6, 119.5, 145.2, 146.6, 164.7, 168.2; HRMS calcd for
C16H22N2O3 291.1709, found 291.1713 (MH+). Compound 13
(a mixture of four diastereomers; two peaks on HPLC): MS
(EI) m/ z 395 (M+).
P r ep a r a tion of [(()-3-[2-(P yr id in -4-yl)eth yl]-2-oxop i-
p er id in -1-yl]a cetic Acid ((()-1). To a 25 wt % solution of
the pyridine-ethyl ester 11 (21.3 g, 73.35 mmol) in isopropyl
alcohol was added 48.8% aqueous sodium hydroxide (7.82 g,
95.36 mmol) at 20 °C under nitrogen over a 5 min period. The
reaction mixture was stirred for 2 h until complete consump-
tion of 11 was observed as monitored by HPLC. The mixture
was cooled to 5-10 °C, seeded with 50 mg of NaCl, and then
quenched by the slow addition of 36.6% aqueous hydrochloric
acid (9.50 g, 95.36 mmol) over
a 10 min period, while
maintaining the internal temperature <15 °C. The final pH
was 5.45. To the resulting mixture was added MeOH (20 mL),
THF (40 mL), and Solka-Floc (5 g). After stirring for 30 min
at ambient temperature, the mixture was filtered through a
pad of Solka-Floc (5 g, wetted with 10 mL of IPA). The filter
cake was washed with a mixture of IPA/THF/MeOH (50 mL:
20 mL:10 mL). The combined filtrate contained acid (()-1 in
quantitative yield as determined by HPLC analysis. The
filtrate was dried by azeotropic distillation under vacuum at
50 °C. After most of the solvents were distilled, the mixture
was flushed several times with IPA (3 × 50 mL) to give a final
concentration of 30 wt % (final weight ) 60 g) and a water
content of <1 mg/mL. The mixture was seeded with (()-1 and
stirred until a seed bed was formed. Hexane (30.5 mL) was
then added over a 1 h period and then aged for 12 h. After
cooling to 10 °C and stirring for 0.5 h, the solid was collected
by filtration. The filter cake was washed with a 40:60 mixture
of IPA:hexanes (50 mL) and vacuum-dried to give 18.3 g of
(()-1 as a light beige crystalline solid. The purity was 96.6
wt %. Thus, the yield was 91.8% (17.7 g) from 11 or an overall
yield of 62% from 7 for the two steps. Complete rejection of
impurities was achieved in this solvent system. The washes
and the mother liquors contained about 10% (2 g) product (()-
1, 15% (3.95 g) bis-pyridyl-acid 14, and 0.9% (0.1 g) piperidone-
acid 8. Pure (()-1: mp 144-145 °C; MS (CI) m/ z 263 (MH+);
1H NMR (CDCl3) δ 1.70 (m, 1H), 1.80-2.05 (m, 4H), 2.20 (m,
1H), 2.40 (m, 1H), 2.78 (t, J ) 8.0 Hz, 2H), 3.35 (m, 1H), 3.47
(m, 1H), 3.90 (A of AB, J ) 17.1 Hz, 1H), 4.32 (B of AB, J )
17.1 Hz, 1H), 7.27 (d, J ) 6.2 Hz, 2H), 8.49 (d, J ) 6.0 Hz,
2H); 13C NMR (CDCl3) δ 17.4, 22.4, 28.1, 28.4, 36.3, 44.9, 45.1,
120.4, 142.7, 149.8, 167.7, 168.3. Anal. Calcd for
C14H18O3N2: C, 64.11; H, 6.92; N, 10.68. Found: C, 64.15; H,
7.16; N, 10.66.
1
70-71 °C; MS (EI) m/ z 185 (M+); H NMR (CDCl3) δ 1.27 (t,
J ) 7.1 Hz, 3 H), 1.85 (br m, 4 H), 2.42 (br m, 2 H), 3.35 (br m,
2 H), 4.10 (s, 2 H), 4.19 (q, J ) 7.1 Hz, 2 H); 13C NMR (CDCl3)
δ 14.1, 21.3, 23.1, 32.1, 48.6, 49.2, 61.1, 169.1, 170.4. Anal.
Calcd for C9H15NO3: C, 58.36; H, 8.16; N, 7.56. Found: C,
58.25; H, 8.51; N, 7.48.
P r ep a r a tion of Eth yl [(()-3-[2-(P yr id in -4-yl)eth yl]-2-
oxop ip er id in -1-yl]a ceta te (11). A 250 mL three-necked
round bottom flask equipped with a stirrer, nitrogen inlet,
cooling unit, and thermometer probe was charged with pip-
eridone-ester 7 (55.6 g, 108.0 mmol; 36 wt % solution from
above), CH3CN (63.0 mL), sodium iodide (17.8 g, 118.8 mmol),
and triethylamine (13.1 g, 129.6 mmol). The mixture was
stirred until all the solid dissolved. The solution was cooled
to 0 °C, chlorotriethylsilane (17.9 g, 118.8 mmol) was added
over 5 min, keeping the internal temperature below +5 °C,
and then the mixture was stirred at 20 °C for 1-2 h. The
resulting mixture was cooled to -5-0 °C, and 4-vinylpyridine
(13.1 g, 124.2 mmol) was added dropwise over a 2 h period,
while keeping the internal temperature below 0 °C. The
reaction was aged at 0 °C for 1-2 h and then quenched by
slow addition into a cold (0 °C) solution of 1 N HCl (140 mL),
while keeping the internal temperature < 20 °C. The final
(23) Kaye, P. T.; Learmonth, R. A. Synth. Commun. 1989, 19, 2337.

Synthesis of Fibrinogen Receptor Antagonist L-734,217
J . Org. Chem., Vol. 61, No. 1, 1996 221
P r ep a r a tion of Qu in in iu m [3(R)-(-)-[2-(P yr id in -4-yl)-
eth yl]-2-oxop ip er id in -1-yl]a ceta te Mon o-2-p r op a n ol Sol-
va te (20). In a 250 mL round bottom flask, pyridine acid (()-1
(12.04 g, 96.6% pure, 44.34 mmol), quinine (14.89 g, 45.90
mmol), and IPA (80.8 mL) were combined. The mixture was
heated at 65 °C for 15 min under a nitrogen atmosphere to
dissolve all the solid. The resulting solution was then allowed
to cool slowly to ambient. When the solution reached 45 °C,
it was seeded with ∼10 mg of quinine salt 20 (99.5% ee). Soon
after the seeding, crystallization began as the temperature
slowly cooled. After stirring overnight at 20 °C, the mixture
was cooled to 5-6 °C and aged for 0.5-1 h. The solid was
collected on a medium porosity fritted funnel under a nitrogen
blanket. The filter cake was washed with 50 mL of cold (5-
10 °C) THF:hexane (50:50) and then dried under vacuum with
a nitrogen sweep to give 12.72 g of 20 as a white solid. The
weight % of the free acid was 42.2%; thus, the yield of free
acid was 46.1% based on 44.34 mmol of (()-1. The optical
purity of the free acid was 98% ee, which was determined by
normal phase HPLC analysis via the chiral amides of
(R)-(+)-1-(1-naphthyl)ethylamine: Column: 4.6 × 250 mm
Zorbax Silica; Eluent: 980:17:3 MeCl2/IPA/15 M aqueous NH4-
OH; flow rate:1.0 mL/min; injection: 20 µL; detection: 260 nm.
Multiple recrystallization gave an analytically pure sample
with >99.5% ee: Compound 20 is a metastable species. Upon
exposure to air, IPA is graduately replaced by water: mp 103-
hydrolysis of IPAC is a problem. An analytically pure sample
of (R)-1 (>99.5% ee) was prepared by crystallization (IPA): mp
128-129 °C; [R]20 ) -21.4 ° (c 1.01, MeOH).
Na
B. Am id e Bon d F or m a tion . To the stirred solution of
the sodium salt of pyridine acid (R)-1 (10.2 mol) from above
was added benzyl 3(R)-aminobutyrate hemisulfate (2500 g,
10.3 mol), IPAC (10.2 L), HOBT (103.2 g, 0.76 mol), and 1-[3-
(dimethylamino)propyl]3-ethylcarbodiimide hydrochloride (2346
g, 12.2 mol). The mixture was stirred at room temperature
overnight. The reaction mixture was extracted with IPAC (2
× 10.2 L). The combined organic layers was washed with
saturated sodium bicarbonate solution (10 L) and then with
water (2 × 20 L). The combined organic solution, contained
∼95% yield of ester 22, which was used directly in the next
reaction. An analytical sample was obtained by concentrating
the organic solution to dryness to give an oily 22: [R]20
)
Na
-1.4° (c 1.03, MeOH); MS (CI) m/ z 437 (M+); 1H NMR (CDCl3)
δ 1.19 (d, J ) 6.7 Hz, 3H), 1.54-2.05 (m, 5H), 2.15-2.4 (m,
2H), 2.48 (ABq, J )15.5, 5.5 Hz, 1H), 2.55 (ABq, J )15.5, 5.5
Hz, 1H), 2.69 (m, 2H), 3.32 (m, 2H), 3.81 (d, J ) 15.2 Hz, 1H),
4.03 (d, J ) 15.2 Hz, 1H), 4.33 (m, 1H), 5.07 (ABq, J ) 13.9,
12.3 Hz, 2H), 6.76 (br d, J ) 8.5 Hz, 1H), 7.13 (d, J ) 6.0 Hz,
2H), 7.32 (s, 5H), 8.46 (d, J ) 6.0 Hz, 2H); 13C NMR (CDCl3)
δ 20.1, 21.7, 26.5, 32.4, 32.6, 40.0, 40.9, 42.0, 49.5, 51.7, 66.3,
123.9, 128.2, 128.3, 128.6, 135.6, 149.7, 150.9, 167.9, 171.3,
173.0; HRMS calcd for C25H32N3O4 438.2393, found 438.2367
(MH+).
105 °C; [R]22 ) -115.5° (c 1.1, MeOH). Anal. Calcd for
Na
C34H42N4O5‚1.5H2O‚0.2C3H8O: C, 66.41; H, 7.51; N, 8.95.
Found: C, 66.52; H, 7.24; N, 8.75.
P r ep a r a tion of N-[[3(R)-(-)-[2-(P ip er id in -4-yl)eth yl]-
2-o x o p i p e r i d i n -1-y l]a c e t y l]-3(R )-m e t h y l-â-a la n i n e
(L-734,217). The pyridine amide benzyl ester 22 solution in
IPAC from the last step was concentrated under vacuum (e40
°C pot temperature) to a volume of 8 L, and then 10 L of MeOH
was added and the solution concentrated again to 8 L. The
flush (temperature e50 °C, 10 cmHg) was repeated four times
until all the IPAC was replaced with MeOH (maximum IPAC
content ) 50 mol % relative to benzyl ester). The resulting
solution was divided into two equal portions and each sub-
jected to the following hydrogenation conditions.
Recyclin g of th e Resolu tion Moth er Liqu or . P r ep a r a -
tion of Qu in in iu m [3(R)-(-)-[2-(P yr id in -4-yl)eth yl]-2-oxo-
p ip er id in -1-yl]a ceta te (20). Ra cem iza tion a n d Resolu -
tion . A 3 L three-necked round bottom flask equipped with a
mechanical stirrer, condenser, nitrogen gas inlet, and ther-
mometer probe was charged with pyridine acid quinine salt
mother liquor from the resolution step (1.84 L, 77.7 g/L, 545
mmol) and solid sodium ethoxide (67 g, 981 mmol). The
mixture was heated at reflux (83-85 °C) for 5 h. The resulting
mixture was cooled to 15 °C, and 37% hydrochloric acid (56
mL) was slowly added until pH 8.1-8.3 (at 23 °C). To the
batch was added THF (1.84 L) and Solka-Floc (80 g). After
stirring for 1 h, the mixture was filtered through a a layer of
Solka-Floc (40 g, prewetted with THF) in a sintered glass
funnel (medium porosity) to remove NaCl and then washed
with THF (300 mL). The filtrate was concentrated under
vacuum at 45 °C and flushed with toluene (3 × 300 mL) to
remove water, followed by isopropyl alcohol (3 × 300 mL) to
remove toluene. The water content of the mixture should be
less than 500 mg/mL after the flushes. Isopropyl alcohol (750
mL) was added to the batch to give a final concentration of 3
mL of IPA/gram of the quinine salt. The mixture was heated
at 70-75 °C for 15-30 min under a nitrogen atmosphere to
dissolve all the solid. The resulting solution was allowed to
cool gradually to ambient. When the batch temperature
reached 45 °C, the mixture was seeded with 0.1 g of 99.5% ee
quinine salt 20. After stirring overnight at 22 °C, the mixture
was cooled to 5-10 °C and aged for 0.5 h. The batch was
filtered through a sintered glass funnel, and the wet cake was
washed with cold hexane/THF (1:1, 10 °C, 2 × 350 mL) and
then dried under vacuum under a nitrogen blanket. A 114-
132 g amount of 20 (32-37% yield) was obtained with ∼89.5
wt % purity (9 wt % IPA, 1 wt % NaCl) and 98+% ee.
P r ep a r a tion of N-[[3(R)-(-)-[2-(P yr id in -4-yl)eth yl]-2-
oxop ip er id in -1-yl]a cetyl]-3(R)-m eth yl-â-a la n in e Ben zyl
Ester (22). A. Sa lt Br ea k in g. In a three-necked flask
charged with IPAC (37 L) was added pyridine acid quinine
salt 20 (6711 g, 39.8 wt % of pyridine acid, 10.2 mol). To this
stirred suspension was slowly added 1 N NaOH (10.1 L). The
final pH of the aqueous solution was 9.5. After separation of
two layers, the aqueous layer was extracted with IPAC (8.5
L). The pH of the aqueous layer was adjusted to 10.4 with 1
N NaOH (245 mL) and again extracted with IPAC (1.4 L). The
recovery of the chiral acid (R)-1 in the aqueous layer was
quantitative. Quinine should be <1.5% in the aqueous layer
before proceeding, otherwise another extraction is required.
Alternatively, MTBE should be use in place of IPAC, if
To a 5-gal stirred autoclave was added the pyridine amide
benzyl ester (2231 g, 5.1 mol) solution in MeOH (total volume
was adjusted to 17.6 L) and acetic acid (27.5 mL). To this
solution was added 10% Pd/C (211.5 g). The mixture was
heated to 50 °C and hydrogenated at 40 psi for 20 h. After
the mixture was cooled to room temperature, it was filtered
through ca. 5 in. thick Solka-Floc (1 kg dried in vacuum oven,
prewashed with 4 × 2 L MeOH), and the solid was washed
with MeOH ( 2 × 2.5 L). The assayed amount of L-734,217
was 3325 g (MW 353, 9.4 mol, 92% from pyridine acid quinine
salt). The filtrate was concentrated under vacuum, and the
total volume was adjusted to 15.3 L (12 L MeOH + 3325 g
L-734,217). This solution was heated to reflux under nitrogen,
and CH3CN (20 L) was added while the solution was at reflux.
The solution was seeded with L-734,217 (0.6 g), and CH3CN
(5 L) was added. The mixture was then stirred for 1 h without
heating during which time the temperature dropped from 61
°C to 52 °C. The mixture was stirred at room temperature
overnight and then filtered. The solid was washed with CH3-
CN (6 L). The solid was dried in vacuo (50 °C, 10 cmHg)
overnight to give 3082 g of L-734,217 as a white, fluffy solid
(MW 353.45, 8.72 mol, 85.5% yield based on pyridine acid
quinine salt). HPLC assay showed ca. 203 g of L-734,217 in
the mother liquor. The solid L-734,217 was 99.8% de. No
other impurity peak was observed by HPLC. L-734,217: mp
237-239 °C; [R]25 ) -57.2° (c 1.0, H2O); 1H NMR (CD3OD)
405
δ 1.14 (d, J ) 6.7 Hz, 1H), 1.33 (m, 4H), 1.50-1.90 (m, 5H),
1.93 (m, 4H), 2.29 (dd, J ) 14.3, 6.7 Hz, 1H), 2.38 (dd, J )
14.3, 6.0 Hz, 1H), 2.40 (m, 1H), 2.95 (dt, J ) 12.8, 2.9 Hz, 2H),
3.38 (m, 4H), 3.91 (d, J ) 16.3 Hz, 1H), 4.02 (d, J ) 16.3 Hz,
1H), 4.14 (sextet, J ) 6.7 Hz, 1H); 13C NMR (CD3OD) δ 20.7,
22.5, 27.2, 29.6, 30.0, 30.2, 34.2, 35.0, 42.5, 44.5, 44.7, 45.2,
50.8, 51.7, 169.5, 175.5, 179.5. Anal. Calcd for C18H31N3O4:
C, 61.17; H, 8.84; N, 11.89. Found: C, 61.22; H, 9.00; N, 11.90.
P r ep a r a tion of N-[[3(R)-(-)-[2-(P yr id in -4-yl)eth yl]-2-
oxop ip er id in -1-yl]a cetyl]-3(R)-m eth yl-â-a la n in e (23). A
solution of the benzyl ester 22 (1220 g, 2.79 mol) in MeOH
(11 L) and acetic acid (184 mL) was hydrogenated in a stirred

222 J . Org. Chem., Vol. 61, No. 1, 1996
Chung et al.
autoclave at rt under 40 psi hydrogen over 10% Pd/C (60 g).
The hydrogen uptake was carefully monitored until 1.0 equiv
of hydrogen was consumed (2.0 h), and the reaction was
stopped. The palladium was removed by filtration through
dry Solka-Floc and washed with MeOH (5 L). The filtrate was
concentrated under vacuum to an oil residue and flushed with
anhydrous CH3CN (4 L). The residue started crystallizing at
the end and it was stirred with a mixture of anhydrous CH3-
CN (2.5 L) and anhydrous MTBE (5 L) overnight. The white
solid was collected by filtration and was washed with a 1:2
mixture of CH3CN/MTBE (750 mL) and MTBE (2 × 1 L). It
was dried in vacuum oven at 50 °C yielding 835 g (86%) of 23.
HPLC analysis showed the sample is about 97% pure contain-
ing trace amount of L-734,217 (∼2%). An analytically pure
8.24 (d, J ) 6.1 Hz, 2H); 13C NMR (CD3OD) δ 20.4, 22.5, 27.3,
33.5, 33.6, 41.4, 42.1, 43.9, 50.7, 51.4, 125.8, 138.0, 149.6, 154.7,
169.8, 175.0, 175.1. Anal. Calcd for C18H25N3O4: C, 62.23;
H, 7.25; N, 12.09. Found: C, 61.95; H, 7.10; N, 12.16.
Ack n ow led gm en t. We would like to thank K.
Mistry and T. M. Dowling for analytical support.
Su p p or tin g In for m a tion Ava ila ble: Figure 1, which
contains C-13 NMR chemical shift assignments for 7, 11, 17,
18, bis-silylated 11 (24), A, 4-vinylpyridine, triethylamine, and
silylated triethylamine (2 pages). This material is contained
in libraries on microfiche, immediately follows this article in
the microfilm version of the journal, and can be ordered from
the ACS; see any current masthead page for ordering
information.
sample was obtained by recrystallization from 1:7 MeOH/CH3-
1
CN. 23: mp 142-143 °C; [R]20 ) -14.2° (c 1.0 MeOH); H
Na
NMR (CD3OD) δ 1.22 (d, J ) 6.7 Hz, 3H), 2.25-1.65 (m, 7H),
2.60-2.35 (m, 3H), 2.77 (m, 2H), 3.35 (m, 2H), 3.96 (d, J ) 15
Hz, 1H), 4.03 (d, J ) 15 Hz, 1H), 7.35 (d, J ) 6.1 Hz, 2H),
J O951214F